Keep It Simple: Data-Efficient Learning for Controlling Complex Systems With Simple Models

Power, Thomas; Berenson, Dmitry

doi:10.1109/lra.2021.3056368

Cited by 9 publications

(4 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Especially for the learned models, we can never hope to collect enough data to produce an accurate model in the entire state space (which is infinite dimensional). Thus [29] and [30] have developed methods to reason about the validity of a (learned) model for a given state and action and used these methods to reason about model uncertainty in planning and control.…”

Section: Sensingmentioning

confidence: 99%

Challenges and Outlook in Robotic Manipulation of Deformable Objects

Zhu

Cherubini

Dune

et al. 2022

IEEE Robot. Automat. Mag.

Self Cite

106

View full text Add to dashboard Cite

Deformable object manipulation (DOM) is an emerging research problem in robotics. The ability to manipulate deformable objects endows robots with higher autonomy and promises new applications in the industrial, services, and healthcare sectors. However, compared to rigid object manipulation, the manipulation of deformable objects is considerably more complex, and is still an open research problem. Tackling the challenges in DOM demands breakthroughs in almost all aspects of robotics, namely hardware design, sensing, deformation modeling, planning, and control. In this article, we highlight the main challenges that arise by considering deformation and review recent advances in each sub-field. A particular focus of our paper lies in the discussions of these challenges and proposing promising directions of research.

show abstract

Section: Sensingmentioning

confidence: 99%

Challenges and Outlook in Robotic Manipulation of Deformable Objects

Zhu

Cherubini

Dune

et al. 2022

IEEE Robot. Automat. Mag.

Self Cite

106

View full text Add to dashboard Cite

show abstract

“…Other works are dedicated to learning deformable dynamics using graph neural networks [18] based on extracted key points or the underlying mesh representations [20]. Power et al [22] use simple models as a cheaper data collection way to improve learning efficiency.…”

Section: Introductionmentioning

confidence: 99%

Goal-Conditioned Action Space Reduction for Deformable Object Manipulation

Wang,

Papallas,

Leonetti

et al. 2023

2023 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Planning for deformable object manipulation has been a challenge for a long time in robotics due to its high computational cost. In this work, we propose to reduce this cost by reducing the number of pick points on a deformable object in the action space. We do this by identifying a small number of key particles that are sufficient as pick points to reach a given goal state. We find these key particles through a geometric model simplification process, which finds the minimal geometric model that still enables a good approximation of the original model at the goal state. We present an implementation of this general approach for 1-D linear deformable objects (e.g., ropes) that uses a piece-wise line fitted model, and for 2-D flat deformable objects (e.g., cloth) that uses a mesh simplified model. We conducted simulation experiments on ropes and cloths, which demonstrate the effectiveness of the proposed method. Finally, the planned paths are executed in a real-world setting for two cloth folding tasks.

show abstract

“…Model predictive control (MPC) methods have been widely used in robotics for applications such as autonomous driving [36], bipedal locomotion [5] and manipulation of deformable objects [25]. For nonlinear systems, sampling based approaches for MPC such as the Cross Entropy Method (CEM) and Model Predictive Path Integral Control (MPPI) [15,36] have proven popular due to their ability to handle uncertainty, their minimal assumptions on the dynamics and cost function, and their parallelizable sampling.…”

Section: Introductionmentioning

confidence: 99%

Variational Inference MPC using Normalizing Flows and Out-of-Distribution Projection

Power¹,

Berenson²

2022

Preprint

Self Cite

View full text Add to dashboard Cite

We propose a Model Predictive Control (MPC) method for collision-free navigation that uses amortized variational inference to approximate the distribution of optimal control sequences by training a normalizing flow conditioned on the start, goal and environment. This representation allows us to learn a distribution that accounts for both the dynamics of the robot and complex obstacle geometries. We can then sample from this distribution to produce control sequences which are likely to be both goal-directed and collision-free as part of our proposed FlowMPPI sampling-based MPC method. However, when deploying this method, the robot may encounter an out-of-distribution (OOD) environment, i.e. one which is radically different from those used in training. In such cases, the learned flow cannot be trusted to produce low-cost control sequences. To generalize our method to OOD environments we also present an approach that performs projection on the representation of the environment as part of the MPC process. This projection changes the environment representation to be more in-distribution while also optimizing trajectory quality in the true environment. Our simulation results on a 2D doubleintegrator and a 3D 12DoF underactuated quadrotor suggest that FlowMPPI with projection outperforms state-of-the-art MPC baselines on both in-distribution and OOD environments, including OOD environments generated from real-world data.

show abstract

Keep It Simple: Data-Efficient Learning for Controlling Complex Systems With Simple Models

Cited by 9 publications

References 18 publications

Challenges and Outlook in Robotic Manipulation of Deformable Objects

Challenges and Outlook in Robotic Manipulation of Deformable Objects

Goal-Conditioned Action Space Reduction for Deformable Object Manipulation

Variational Inference MPC using Normalizing Flows and Out-of-Distribution Projection

Contact Info

Product

Resources

About